Hunter Hagewood named as ACCRE’s new Director of Research Computing Operations
Aug. 19, 2019—We are pleased to welcome Hunter Hagewood as ACCRE’s new Director of Research Computing Operations. Hunter takes over from Will French, who left ACCRE earlier this summer. Hunter joins ACCRE after over 10 years in the private sector: first as a founder and CTO of Nevoa Networks, one of the first commercially available cloud storage...
New Python guidelines: ACCRE-optimized Python 3 builds with virtual environments are strongly encouraged
Jul. 15, 2019—We have significantly revised our documentation on using Python on the ACCRE cluster to address some of the most common issues that are brought to our attention and improve the overall cluster user experience. Here is a brief overview of some changes: Although Anaconda provides an easy way to use a large selection of software,...
GPFS briefly unmounted at 8:02am on Wednesday; check output of jobs running at the time
Jul. 24, 2019—At approximately 8:02 AM today a command was issued to remove a misconfigured compute node from the cluster in GPFS. Due to the nature of the configuration problem and the timing of the command, this caused most or all of the ACCRE gateway and compute nodes to unmount GPFS for approximately one minute. We are...
Will French steps down as ACCRE’s Director of Research Computing Operations
Jul. 2, 2019—Will French, ACCRE’s director of research computing operations, has left Vanderbilt after nearly five years of service to ACCRE. Will came to Vanderbilt in 2008 as a graduate student in chemical and biomolecular engineering; he joined us shortly after completing his PhD and eventually became co-director of ACCRE in 2017.
Existing ACCRE groups: Let us help you use ACCRE to your fullest potential.
Jun. 24, 2019—Following best practices in an HPC cluster to ensure job performance and reliability may not be as trivial as it seems, especially with massive workloads, complex pipelines or parallel jobs. ACCRE users and groups can take advantage of our consulting services to optimize their workflow in order to maximize the output of their computations. This...
GPFS data migration to run most weekends, starting tonight
Jun. 7, 2019—Beginning tonight we will kick off a data migration between different pools of storage in our GPFS23 filesystem (/scratch and /data). Hot data (files that have been accessed more recently) will be moved to faster, more expensive storage while cold data (files that have not been accessed recently) will be moved to more economical storage...
Custom/private gateway work is complete
May. 1, 2019—Update, 5/10: The work is now complete. On Friday, May 10 beginning at 8:30AM we will be taking custom/private gateways offline in order to reboot and upgrade the operating system to CentOS 7.6 from 7.4 and the GPFS filesystem to 5.0.2 from 4.2.3. We expect these upgrades to last roughly 1-2 hours. If this time...
4/24/2019: Turbo-Charged Machine Learning: Fitting Models Fast with H2O with Jesse Spencer-Smith
Apr. 22, 2019—Jesse Spencer-Smith, Chief Data Scientist at the Vanderbilt Data Science Institute, presented on April 24, 2019. Due to technical difficulties we couldn’t record this seminar, however materials are available at the following repo on GitHub: https://github.com/vanderbilt-data-science/turbo_h2o Abstract: The model fitting stage of a data science workflow has always been compute-intensive. Combine this with a need...
GPFS outage resolved; check output of running jobs
Apr. 19, 2019—Around 2PM today a GPFS manager node had an issue and caused the GPFS filesystem’s to hang across the cluster, making logins and file access unresponsive. The issue was corrected at 3PM today and all compute nodes seem to have recovered. Please check the output of running jobs just to be safe.
Scheduled maintenance on public gateways complete
Apr. 18, 2019—Original post: This Saturday morning, April 20th, we will be taking the public gateway and portal servers offline from 7 am to 9 am in order to reboot and upgrade the operating system to CentOS 7.6 from 7.4 and the GPFS filesystem to 5.0.2 from 4.2.3. Updating the entire cluster to GPFS 5 is an...